Annals of Surgical Oncology

, Volume 17, Issue 9, pp 2264–2273

Body Mass Index and Major Cancer Surgery Outcomes: Lack of Association or Need for Alternative Measurements of Obesity?


    • Department of SurgeryUniversity of Minnesota and Minneapolis VAMC
  • Helen M. Parsons
    • Division of Health Policy and ManagementUniversity of Minnesota
  • William G. Henderson
    • The Colorado Health Outcomes ProgramThe University of Colorado
  • Eric H. Jensen
    • Department of SurgeryUniversity of Minnesota
  • Todd M. Tuttle
    • Department of SurgeryUniversity of Minnesota
  • David A. Rothenberger
    • Department of SurgeryUniversity of Minnesota
  • Todd A. Kellogg
    • Department of SurgeryUniversity of Minnesota
  • Beth A. Virnig
    • Division of Health Policy and ManagementUniversity of Minnesota
Healthcare Policy and Outcomes

DOI: 10.1245/s10434-010-1023-2

Cite this article as:
Al-Refaie, W.B., Parsons, H.M., Henderson, W.G. et al. Ann Surg Oncol (2010) 17: 2264. doi:10.1245/s10434-010-1023-2



Although surgical studies have reported inconsistent associations between increased body mass index (BMI) and operative outcomes, the accuracy of BMI for measuring obesity has been questioned in previous epidemiologic studies. Simultaneously, BMI has known comorbidities, which may mediate the effect of BMI if included in multivariable models. We sought to examine the effect of BMI on operative outcomes after adjusting for preoperative factors.


We identified 8858 patients who underwent major thoracic, abdominal, and pelvic surgery for solid organ tumors in American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) centers from 2005 to 2007. We used multivariable analyses to assess the effect of BMI on short-term operative outcomes after controlling for covariates.


Increased BMI was not associated with worse short-term operative outcomes in our bivariable analyses. However, patients with BMI ≥35 had higher American Society of Anesthesiologists scores, longer operative times, and an increased number of postoperative complications (P < 0.0001). After adjusting for pre- and intraoperative factors, BMI did not predict any short-term operative outcome except for an increased total number of complications in BMI ≥35. These results persisted after removing potential mediators from the multivariable analysis.


In ACS NSQIP, BMI has minimal association with short-term operative outcomes after major cancer surgery. Although these findings may suggest a lack of association between obesity and cancer surgery outcomes, it confirms the previously examined limitations of BMI. Because of the rising incidence of obesity in the United States and its challenging effect on surgeon’s practice, ACS NSQIP should consider exploring alternative measures of general and abdominal obesity.

Copyright information

© Society of Surgical Oncology 2010